Several Useful Databases

The following online resources provide  a lot of useful databases that can be set as benchmarking datasets for machine learning tasks (regression, classification, cluster, etc.) . Please note that some information may occur repeatly in the list.

 

       Widely-used Ones:

• The Mnist Database for Handwritten Digits Recognition  http://yann.lecun.com/exdb/mnist/
• LIBSVM http://www.csie.ntu.edu.tw/~cjlin/libsvmtools/datasets/multiclass.html#usps
• UCI Machine Learning Repository http://archive.ics.uci.edu/ml/index.html
• Delve Datasets http://www.cs.toronto.edu/~delve/data/summaryTable.html
• StatLib-Datasets Archive http://lib.stat.cmu.edu/datasets/
• Feature Selection Datasets http://featureselection.asu.edu/datasets.php
• Virtual Library of Simulation Functions  http://www.sfu.ca/~ssurjano/index.html
• KEEL http://www.keel.es/
• Indian Script Character Databases:Link
• Economic Research: https://fred.stlouisfed.org/series/DEXUSEU
  

  Miscellaneous Ones

• Cambridge Handwritten Word Imageshttp://documents.cfar.umd.edu/resources/database/handwriting.database.html
• CEDAR Handwritten Database http://www.cedar.buffalo.edu/Databases/
• Signature Verification Competition 2004 Database http://www.cs.ust.hk/svc2004/download.html/
• MCYT Online and Offline Signature Database http://atvs.ii.uam.es/bbdd_EN.html
• Caltech Signature Database http://www.vision.caltech.edu/mariomu/research.html
• Offline GPDS signature database http://www.gpds.ulpgc.es/download/
• HIT-MW Chinese Signature database http://hitmwdb.googlepages.com/writeridentification
• LIBSVM Data: Classification, Regression, and Multi-label [Link]

 

        To be Continued…